基于BN网络的学生广义认知诊断模型研究

Research on BN-based generalized cognitive diagnosis model for students

针对传统教学中对学生心理变化和广义认知的测量与诊断只给出分数值,而无法对分数值相同的不同知识结构进行解释的缺点,文中基于目前学生认知诊断的任务更多的是学生测试结果的反馈信息,将贝叶斯网络(BN)引入学生广义认知诊断中。主要进行了两个研究,分别为测验所得数据进行贝叶斯结构的学习得到属性间的层级关系与构建BN网络分类器对学生认知状态进行分类。最后,对所构造的基于BN网络的学生广义认知诊断模型进行验证。结果表明,其得到的属性层级关系合理,分类性能良好,具有广阔的应用前景。

In allusion to the disadvantage that only score values are given for measurement and diagnosis of students′psy-chological change and generalized cognition,and different knowledge structures with the same score value cannot be explained in the traditional teaching,the Bayesian network(BN)is introduced into the generalized cognitive diagnosis of students on the basis of the fact that,as for the current task of student cognitive diagnosis,information feedback of student test results is more important.The Bayesian structure learning is conducted for the data obtained from the test,so as to obtain hierarchical relation-ships between attributes.The BN network classifier is constructed to classify the students′cognitive states.A verification was conducted for the constructed BN-based generalized cognitive diagnosis model of students.The results show that the obtained hierarchical relationships between attributes are reasonable,the classification performance is good,and the model has a broad application prospect.